MMest_loccov: S- and MM-Estimates of multivariate location and covariance...

In FRB: Fast and Robust Bootstrap

S- and MM-Estimates of multivariate location and covariance matrix

Description

Compute S- and MM-Estimates of multivariate location and covariance matrix

Usage

    MMest_loccov(Y, control=MMcontrol(...), ...)
    Sest_loccov(Y, bdp=.5, control=Scontrol(...), ...)
    MMest_twosample(X, groups, control=MMcontrol(...), ...)
    Sest_twosample(X, groups, bdp=0.5, control=Scontrol(...), ...)

Arguments

Y	input matrix or data frame
X	input matrix or data frame
bdp	breakdown point, defaults to 0.5
groups	grouping variable
control	a list with control parameters for tuning the S- or MM-estimate and its computing algorithm, seeScontrol and MMcontrol.
...	further arguments to be passed to CovMMest()

Details

This functions are internal, wrappers around the functions Sest() CovMMest().

Value

Return lists with the following components:

Mu	location
Gamma	shape
scale	scale=det^(1/(2*m))
Sigma	covariance matrix
c1	tuning parameter of the loss function for MM-estimation
SMu	location of the initial S-estimate
SGamma	shape of the initial S-estimate
SSigma	covariance matrix of the initial S-estimate
b	tuning parameters used in Tukey biweight loss function for S-estimation, as determined by bdp
w	scaled weights
outflag	outlier flags

Examples

    Y <- matrix(rnorm(50*5), ncol=5)
    (MMests <- MMest_loccov(Y)) 

    (Sests <- Sest_loccov(Y, bdp = 0.25)) 

    Y1 <- matrix(rnorm(50*5), ncol=5)
    Y2 <- matrix(rnorm(50*5), ncol=5)
    Ybig <- rbind(Y1,Y2)
    grp <- c(rep(1,50),rep(2,50))
    (MMests <- MMest_twosample(Ybig, grp))
    

FRB documentation built on Oct. 7, 2024, 5:09 p.m.